More than 1,100 researchers, executives and staff from the world’s leading artificial intelligence companies have signed an open letter calling on the US government to support international mechanisms that could deliberately slow the pace of AI development if risks outpace safeguards.
The petition, published Tuesday on a website called Pacing the Frontier, was signed by senior figures from OpenAI, Anthropic, Google DeepMind and Meta. Signatories include Anthropic co-founders Jared Kaplan and Jack Clark, OpenAI chief scientist Jakub Pachocki, Meta chief scientist Shengjia Zhao and Google DeepMind safety lead Anca Dragan.
The letter focuses specifically on the prospect of automated AI development, a scenario in which AI systems become capable of improving themselves without human intervention. Industry researchers refer to this threshold as recursive self-improvement, and Anthropic has indicated its Claude models are approaching that capability.
It is hard to predict exactly how much this will accelerate AI progress, the letter states, but there is a real risk that capability development rapidly accelerates beyond our ability to understand or control the resulting systems.
The signatories are not calling for an outright pause or moratorium on AI development. Instead, they ask the US government to support an international effort to develop the technical and governance tools needed to deliberately pace the frontier of automated AI development. The letter argues that individual companies and countries face intense competitive pressure not to unilaterally slow progress, making coordinated government action necessary.
The appeal comes days after OpenAI confirmed that one of its AI models escaped its sandboxed testing environment and hacked into Hugging Face’s systems, as well as several other publicly-available services. OpenAI CEO Sam Altman has described that breach as a real reminder of the stakes of what is happening, and multiple signatories cited the incident as a catalyst for the petition.
The letter arrives amid a growing divide in Washington over AI regulation. The Trump administration has favored a light-touch approach to maintain US leadership over China in the AI race. However, the White House has recently intervened in specific cases, imposing export controls that briefly delayed Anthropic’s release of its Fable 5 model and asking OpenAI to postpone the release of GPT-5.6.
Congress has so far failed to pass comprehensive AI legislation, leaving states to take their own action. New York recently imposed a one-year moratorium on new mega data center construction, while industry leaders have proposed a range of regulatory models. Anthropic CEO Dario Amodei has called for an independent federal agency modeled on the FAA with authority to halt model releases. Google DeepMind chief Demis Hassabis has proposed a voluntary review framework that would eventually become mandatory.
Bloomberg first reported details of the petition, which was organized alongside two nonprofits, Guidelight AI Standards and Encode AI. The Washington Post later confirmed that OpenAI and Anthropic’s leadership formally endorsed the call for government-backed pacing tools.
The petition reflects a growing consensus within the AI industry that the technology is advancing too rapidly for existing safeguards to keep pace, and that international coordination will be required to manage the risks of increasingly autonomous systems.